This paper presents a text line segmentation method for printed or handwritten historical Arabic documents. Documents are first classified into 2 classes using a K-means scheme. These classes correspond to document complexity (easy or not eas

光学字符识别（即OCR）技术。给出了一些基于linux/KDE的英文OCR代码，该代码对于如何实现一个OCR系统有很好的启发性。-Optical character recognition (ie OCR) technology. Gives some based on linux/KDE English OCR code, the code for how to achieve an OCR system is a good instructive.